Awais Belal [Tue, 14 Feb 2017 21:14:15 +0000 (16:14 -0500)]
kernel-yocto.bbclass: ensure repatching when HEAD is checked out
In some cases it is seen that kernel_checkout and validate_branches are
run again in simultaneous builds. During do_patch the kgit-s2q mechanism
looks for a sentinel file inside the .git directory, finds a fence post
and starts picking up patches after that.
This can create trouble as validate_branches checks out the HEAD of the
branch and so the patches should be reapplied rather than skipped due to
finding of the fence post.
We can call kgit-s2q --clean to remove the sentinel file when the
branches are checked out.
Signed-off-by: Awais Belal <awais_belal@mentor.com> Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ed Bartosh [Fri, 10 Feb 2017 10:41:15 +0000 (12:41 +0200)]
wic: direct: move UUID generation to PartitionedImage
Moved code that generates partition UUIDs from DirectPlugin to
PartitionedImage class as it's more logical to have it there.
This allows the code to be reused by other imager plugins.
Got rid of having yet another list of partitions in PartitionedImage.
Reused the list passed from DirectPlugin.
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Ed Bartosh [Fri, 10 Feb 2017 10:10:22 +0000 (12:10 +0200)]
wic: direct: add 'realnum' attribute to partition
Replaced call of _get_part_num method with an attribute.
This eliminates the need to call the method and loop over
partitions every time we need to know realnum for partition.
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Ed Bartosh [Thu, 9 Feb 2017 13:52:55 +0000 (15:52 +0200)]
wic: move disk operations to PartitionImage class
Disk operations were spread over DirectPlugin, DiskImage and Image
code making the code hard to understand.
Renamed Image class to PartitionedImage.
Removed DiskImage class.
Moved disk operations to PartitionedImage.
There was an implicit support for multiple disks: if different devices
were specified in .wks file (e.g. --ondisk sda and --ondisk sdb), wic
would theoretically generate multiple images. This is quite confusing
option and the code supporting it was broken for a long time. The same
effect (multiple output images) can be achieved in obvious and clear
way - by using multiple .wks files.
This functionality was removed. PartitionedImage works only with
one image. This makes the code less complex and easier to maintain.
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Kristian Amlie [Mon, 6 Feb 2017 16:16:46 +0000 (17:16 +0100)]
wic: Add --exclude-path option to rootfs source plugin.
It will omit the given path from the resulting partition, and if the
given path ends in a slash, it will only delete the content, and keep
the directory.
Since mkfs only accepts whole directories as input, we need to copy
the rootfs directory to the workdir so that we can selectively delete
files from it.
Since we want to use the copyhardlinktree() function, we need to put
the generic oe lib in the module search path.
Signed-off-by: Kristian Amlie <kristian.amlie@mender.io> Signed-off-by: Ross Burton <ross.burton@intel.com>
Robert Yang [Fri, 10 Feb 2017 03:23:13 +0000 (19:23 -0800)]
strace: fix gawk's path
Fixed:
strace-4.15: /usr/lib/strace/ptest/tests/caps-abbrev.awk contained in package strace-ptest requires /bin/gawk, but no providers found in RDEPENDS_strace-ptest? [file-rdeps]
The path should be /usr/bin/gawk as other scripts use in this package.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Recent kernel build jvmti, java profiling agent, if a host jdk is
detected. The library will be packaged in -dev and throws a qa error.
Disable the jvmti build per default.
Signed-off-by: Stefan Müller-Klieser <s.mueller-klieser@phytec.de> Signed-off-by: Ross Burton <ross.burton@intel.com>
Hongxu Jia [Wed, 8 Feb 2017 03:01:35 +0000 (22:01 -0500)]
elfutils: update homepage and upstream souce
In git://sourceware.org/git/elfutils.git:
----------------
commit 1700fd25e6caf26663af2bd994d1d99fab9df59f
Author: Mark Wielaard <mark@klomp.org>
Date: Sat Dec 24 22:31:41 2016 +0100
http://elfutils.org/ is now hosted at http://sourceware.org/elfutils/
fedorahosted used to be our home, but we are now hosted at sourceware.
Change the elfutils project home to http://elfutils.org/
Point hosted services (email, release, git, bug tracker and web pages)
to https://sourceware.org/elfutils/
Move design notes from README to NOTES.
Add URLs for home, releases, bugs, git and mailinglist to README.
Make the --version output of all tools the same by using a common
print_version function and update the publicly shown copyright holder
to the elfutils developers.
Signed-off-by: Mark Wielaard <mark@klomp.org>
------------------
Hongxu Jia [Tue, 7 Feb 2017 03:29:07 +0000 (22:29 -0500)]
gpgme: 1.6.0 -> 1.8.0
- Drop gpgme.pc, it is duplicated with pkgconfig.patch
- Rebase pkgconfig.patch to 1.8.0
- Split cmake config to ${PN}-dev
- Update LIC_FILES_CHKSUM, because copytright year changed.
- The libgpgme-pthread is obsolete by upstream;
- Add package libgpgcpp;
TODO: Do not remove gpgme-pthread.pc from pkgconfig.patch,
the upstream explains:
git://git.gnupg.org/gpgme.git
commit 09b64554328445e99a8cc78fc34ea49c2ea2e7f9
------------------------------
libgpgme-pthread is removed but gpgme-config still supports
--thread=pthread for compatibility with find scripts.
------------------------------
Fabio Berton [Wed, 25 Jan 2017 11:22:05 +0000 (09:22 -0200)]
watchdog: Add RCONFLICTS to watchdog-keepalive and split RDEPENDS
watchdog-keepalive is a simplified version of the watchdog daemon
and can't be installed along with watchdog daemon, we need to
choose which daemon install.
Signed-off-by: Fabio Berton <fabio.berton@ossystems.com.br> Signed-off-by: Ross Burton <ross.burton@intel.com>
Chen Qi [Wed, 8 Feb 2017 09:25:36 +0000 (17:25 +0800)]
console-tools: fix string formatting issue
Fix string formatting issue in console-tools. Otherwise, when compiling with
'-Wformat -Wformat-security -Werror=format-security', we would have compilation
error.
[YOCTO #9540]
Signed-off-by: Chen Qi <Qi.Chen@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
He Zhe [Wed, 8 Feb 2017 02:58:46 +0000 (10:58 +0800)]
ltp: Fix containers/userns05 failure for lib32
Backport a patch to fix userns05 case:
<<<test_output>>>
user_namespace5 1 TFAIL : userns05.c:95: userns:parent should be not equal to cpid2
user_namespace5 1 TFAIL : userns05.c:95: userns:parent should be not equal to cpid2
user_namespace5 0 TINFO : Child process returned TPASS
user_namespace5 0 TINFO : Child process returned TPASS
user_namespace5 0 TINFO : Child process returned TFAIL
incrementing stop
Rename the previous patch to add a number prefix
Signed-off-by: He Zhe <zhe.he@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Backport a patch from 2.7 branch to fix a regression with glibc
2.24 causing "OSError: [Errno 38] Function not implemented" when
calling urandom() with older kernels.
Signed-off-by: Andreas Oberritter <obi@opendreambox.org> Signed-off-by: Ross Burton <ross.burton@intel.com>
Herve Jourdain [Thu, 9 Jun 2016 07:41:31 +0000 (15:41 +0800)]
smartpm: if RPM transaction fails with an empty list of problems in attempt mode, ignore it!
Since commit be0cabf816a84c59e0e951b92a570121ed843822 on 2016/05/17, smartpm will retry the transaction if ts.run() returns with an empty list of problems, and we're in attempt mode.
But then, since all the packages are already installed, the transaction will fail with "package already installed" error.
In attempt mode, it's fine to not retry nor generate an error when the transaction fails for no reason linked to files/packages conflicts, so this patch fixes this behaviour.
Signed-off-by: Herve Jourdain <herve.jourdain@neuf.fr> Signed-off-by: Ross Burton <ross.burton@intel.com>
The checksum of the LICENSE file changed due to some clarifications.
There were no changes to the actual licensing terms.
The LICENSE variable was not accurate, so I made changes to it.
Specifically:
* there's no GPL code in PulseAudio so I dropped GPL from the list
* the LGPL code allows using later versions of the license rather than
limiting to just 2.1
* there are some MIT and BSD licensed bits
I added more files to LIC_FILES_CHKSUM to have better coverage of all
the differently licensed code.
Dropped json-c and gdbm from DEPENDS. The new release doesn't use json-c
any more. gdbm isn't used when --with-database=simple is passed to
configure, so it should have been removed from DEPENDS a long time ago.
The new release dropped the Xen module, so the --without-xen configure
option isn't needed any more.
Added a comment for why --without-fftw is used.
Disabled the adrian echo canceller, because it has an unusual license,
and disabling the code was simpler than adding a new license to OE-Core.
Dropped upstreamed patches.
Signed-off-by: Ross Burton <ross.burton@intel.com>
because of the way that the test cases are looked in the results,
the file logger is incompatible with extended implementations of the
unittest runner.
as the xml runner extends the unittest runner, it shares the id()
method which returns the full name of the test, not only the test
function name. With that, a single check of the full name reviews at
the same time the class name as well as the function name.
[YOCTO#11012]
Signed-off-by: Benjamin Esquivel <benjamin.esquivel@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Markus Lehtonen [Fri, 3 Feb 2017 17:58:54 +0000 (19:58 +0200)]
oe-build-perf-test: drop --commit-results
Drop support for committing results into Git repository. The
functionality was not directly related to testing and feels unnecessary
complication of the script. The functionality has been moved into a
separate oe-git-archive script.
[YOCTO #10582]
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Markus Lehtonen [Thu, 2 Feb 2017 14:30:57 +0000 (16:30 +0200)]
scripts: Implement oe-git-archive
A helper script for committing data to git and pushing it upstream. The
motivation for the script stems from the need to archiving QA test
results in git, but it could be used elsewhere, of course. The script
needs to be run under an initialized build environment because it
utilizes bitbake configuration information.
[YOCTO #10582]
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Markus Lehtonen [Fri, 27 Jan 2017 14:54:53 +0000 (16:54 +0200)]
oeqa/buildperf: don't archive stdout/stderr of commands
Stop capturing output of the shell commands into <test>/commands.log.
Redirecting output into a file prevented the unittest framework from
capturing it, causing useless errors (with empty output) like:
oeqa.utils.CommandError: Command '['bitbake', 'core-image-sato']'
returned non-zero exit status 1 with output:
In general, the console output of commands is only interesting when
something fails. Also, dropping the commands.log file is a huge saving
in disk space, and thus, repository size when results are archived in
Git.
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Joshua Watt [Sat, 11 Feb 2017 03:35:37 +0000 (21:35 -0600)]
classes/license: Respect beginline and endline
Recipes have the option of specifying a 'beginline' and 'endline'
options in LIC_FILES_CHKSUM which will cause the license bbclass to
only extract those specific lines from the license file when generating
the license database.
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Amarnath Valluri [Fri, 10 Feb 2017 13:42:17 +0000 (15:42 +0200)]
package_deb.bbclass: Ignore file paths in RPROVIDES
Unlike rpm, debian packaging does not allow file paths in 'Provides:' field.
When 'usrmerge' distro feature enabled bash/busybox packages adds '/bin/sh' to
it's RPROVIDES to satisfy build dependencies, this entry should be filtered out.
Signed-off-by: Amarnath Valluri <amarnath.valluri@intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Mariano Lopez [Thu, 9 Feb 2017 13:15:14 +0000 (13:15 +0000)]
testimage.bbclass: Add check for empty test suite
If for some reason (most likely incorrect TEST_SUITES var)
you end with an empty test suite, testimage will start the
target (qemu by default) and will run an empty suite, so
better have a check before starting the target.
[YOCTO #10979]
Signed-off-by: Mariano Lopez <mariano.lopez@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Nathan Rossi [Mon, 30 Jan 2017 07:39:12 +0000 (17:39 +1000)]
image.bbclass: Use 'populate_sdk_base' for non-linux targets
When 'populate_sdk_ext' was first introduced in commit bf81d6bb7f6 it
replaced the inheriting of 'populate_sdk_base'. For non-linux targets
building the extensible SDK caused build errors, and the image class was
changed to inherit 'populate_sdk' when targeting a non-linux SDK_OS (in
commmit e471ce3464d). However inheriting 'populate_sdk' instead of
'populate_sdk_base' causes the SDK to always be built, this is not
expected for the image class.
This change makes the image class inherit 'populate_sdk_base' in the
non-linux SDK_OS case so that it behaves the same as it is expected to
behave where 'bitbake <image> -c populate_sdk' must be executed to
generate the SDK deployables.
Signed-off-by: Nathan Rossi <nathan@nathanrossi.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Philip Balister [Wed, 8 Feb 2017 20:27:06 +0000 (15:27 -0500)]
sanity.bbclass: Add which to SANITY_REQUIRED_UTILITIES.
Using docker with the Fedora 23 container exposed an issue with
the rpm-native configure step. If which is not present the configure
script fork bombs. After much pain, I tracked this to which not being
present in the default container. Add a check for which so others
do not have to have this experience.
Signed-off-by: Philip Balister <philip@balister.org> Signed-off-by: Ross Burton <ross.burton@intel.com>
Paul Eggleton [Tue, 7 Feb 2017 19:12:29 +0000 (08:12 +1300)]
classes/copyleft_filter: include AGPL in default included list
If we're including all GPL licenses then we should include AGPL in
the default COPYLEFT_LICENSE_INCLUDE since it has similar terms (and
currently we do have one AGPL-licensed recipe in OE-Core, namely db).
(Note that the default list isn't meant to be a guaranteed complete
safe list - seek legal advice based on what you're distributing if
unsure.)
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Andrea Adami [Wed, 25 Jan 2017 23:56:08 +0000 (00:56 +0100)]
kernel.bbclass: Make symbol to vmlinuz.bin in boot directory
The changes committed with 6e58f54be103814b6b8a85b236510633c49e6832 did only
consider the vmlinuz ELF kernel but some devices do reqire a raw vmlinuz.bin
(as for example gcw0 in meta-handheld booted by ubiboot).
Expand the above mentioned patch to consider this case.
Signed-off-by: Andrea Adami <andrea.adami@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Jussi Kukkonen [Thu, 9 Feb 2017 19:38:33 +0000 (21:38 +0200)]
cve-check-tool: Use CA cert bundle in correct sysroot
Native libcurl looks for CA certs in the wrong place by
default.
* Add patch that allows overriding the default CA certificate
location. Patch is originally from meta-security-isafw.
* Use the new --cacert to set the correct CA bundle path
Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Jussi Kukkonen [Thu, 9 Feb 2017 19:38:32 +0000 (21:38 +0200)]
cve-check-tool: Fixes for recipe sysroots
* Use --enable-relative-plugins so cve-check-tool looks for
loadable modules relative to binary location instead of
hard-coding a wrong sysroot location
* do_populate_cve_db() assumes that the binary cve-check-update is in
the sysroot. Ensure that this is true by adding a task dependency
Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Jussi Kukkonen [Thu, 9 Feb 2017 19:38:31 +0000 (21:38 +0200)]
cve-check.bbclass: Fix dependencies
With recipe-specific sysroots the cve_check task must depend on
cve-check-tool-native:do_populate_sysroot to get the cve-check-tool
binary into the recipe sysroot.
A normal DEPENDS isn't used to avoid cyclic dependencies.
Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Juro Bystricky [Mon, 6 Feb 2017 17:49:37 +0000 (09:49 -0800)]
testimage: Implement custom target loading
This patch implements custom target loading for testimage, currently
missing due to major changes to the test framework.
Custom targets can be defined in various meta-layers, so we
need an extra path information in order to find them.
Any other additional info is retrieved as usual via the variables
TEST_TARGET and TEST_SUITES
Signed-off-by: Juro Bystricky <juro.bystricky@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 9 Feb 2017 17:11:38 +0000 (17:11 +0000)]
classes: Drop now unneeded update_data calls
Now that the datastore works dynamically we don't need the update_data calls
so we can just remove them. They're not actually done anything at all for
a while.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Wed, 8 Feb 2017 16:17:09 +0000 (16:17 +0000)]
sstate: Make absolute symlinks an error
The current relocation code is broken, at least in the native case. Fixing it
would mean trying pass in new data on sstate tasks about the relative positioning
of symlinks compared to the sstate relocation paths. Whilst we could do this,
right now I'm favouring making this an error and fixing the small number of
problematic recipes we have in OE-Core (3).
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Wed, 8 Feb 2017 16:10:50 +0000 (16:10 +0000)]
openssl/fontconfig/bzip2: Use relative symlinks instead of absolute ones (using a new class)
Absolute path symlinks are a bit of a pain for sstate and the native versions
of these recipes currently contain broken symlinks as a result. There are
only a small number of problematic recipes, at least in OE-Core, namely the
three here.
Rather than trying to make sstate handle this magically, which turns out to
be a harder problem than you'd first realise, simply make the symlinks relative
early in the process and avoid all the problems.
The alternative is adding new complexity to sstate which we could really
do without as without the complexity, you can't always tell where the
absolute symlink is relative to (due to prefixes used for native sstate).
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We used to need to build gcc-cross-initial against a bare sysroot to avoid
contamination. With RSS, we no longer need to do this since the recipe sysroot
is already bare. We can therefore simply point at that and drop this code.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
